How to search inside videos? and more importantly; why would your enterprise need smart search to search inside videos in the first place?
The answer is simple! Video use is rising every day. Take Zoom for instance, where organizations are now recording online meetings and storing them. As video use rises, it becomes difficult to find that one video amongst a thousand others.
With search inside video, you can speed up the process of looking for the video that you need. You can search for spoken words, faces of people and much more. But how does this work? Learn more in this blog.
The Concept Behind Smart Search
To understand how smart search inside video works, it is important to first understand the concept behind it. This section helps you understand this concept by first explaining how normal search engines work and then expanding the concept on to smart search.
A normal search engine can help you search inside documents. When you type a word in the search bar, the search engine checks every document in the repository, and it returns the document with the greatest number of words resembling the word that you just typed in. If you type in the word “Compliance Training 2019”, the document that has word “Compliance Training 2019” the most in its text will show up first.
In smart video search, a similar thing happens. Smart search looks through the transcript of the video for the word that you type in. It returns the video, which has the spoken word repeated the most number of times. If you type in the word “Compliance Training 2019”, the video that has word “Compliance Training 2019” spoken the most, will show up first.
Search can be further enhanced through metadata. This is a description that is attached to the video and it helps identify the content of the video. Examples of such descriptions can include the faces in the video, objects, location, or even tags. This way, if you add the description "Dr. John appears in this video" to a video, then that video will appear when you type in "Dr. John" in the search bar.
AI helps further improve this process, by automatically adding metadata to videos, every time they are uploaded. This way, users won't have to tag videos manually. In the above example, "Dr.John" can be automatically detected through facial recognition, and users won't have to go through every video and tag "Dr.John"
The Tools You Need for Smart Video Search
There are multiple ways in which you can search inside videos; through your desktop application, through your content management system (eg. SharePoint), or through a video content management system. We will walk you through each:
For videos stored on your employees' PCs, you can optimize them through Windows features. Windows allows you to add metadata to your video files and this can be done by following these steps:
- Navigate to your video in Windows Explorer.
- Right Click and Open properties.
- Open the details tab.
You can now add metadata on the following screen:
This method isn't the best, as you can only add limited metadata. More importantly, you can't search through spoken words. Lastly, most data in enterprises is on the cloud or on an internal network and you'd need a content management system to search inside videos for this.
Content Management Systems
Most content management systems allow you to add metadata to video files, in a similar way as it is done on your desktop. You can navigate to your video file and add tags to the file's properties.
Content management systems allow you to add customized fields in the properties of a file. This can be done through the application or by contacting your vendor; depending on which CMS you use. The issue, however, is that you can't search inside videos for spoken words, faces or objects and for this, you would need a video content management system.
Video Content Management Systems
These are cloud video platforms that allow you to host videos and efficiently manage such videos. Examples of such platforms include Qumu, Brightcove, Kaltura, Panopto and VIDIZMO.
Depending on the platform, you get multiple smart video features that allow you to easily look for that one video amongst thousands.
Here are 11 ways in which VIDIZMO can help you search inside videos:
- Search by words spoken in a video: VIDIZMO offers automatic speech recognition (ASR) software to provide machine transcription, which helps index every spoken word in the video, thereby allowing users to search for any keywords or phrases spoken in a video. This way, users can quickly skip to any precise portion of the video where the searched keyword is used.
- Search by on-screen text: Using its automatic speech recognition (ASR) software, VIDIZMO also generates closed captions or video subtitles that display the text version of spoken words on the video screen. The text in the closed captions is indexed so you can search by words or on-screen subtitles that appear in a video.
- Search by manual transcript or existing closed captions: VIDIZMO also supports both the incorporation and indexing of manually generated human transcripts and existing closed caption files to improve video search. Additionally, for both manual and automatically generated machine transcripts, VIDIZMO allows speedy editing of the transcribed text directly from the transcription pane to achieve greater accuracy of the transcribed text.
- Search by the speaker: VIDIZMO utilizes intelligent facial detection software for speaker recognition within videos. Different sections of the video are marked based on the person speaking at various intervals so you can quickly jump to a particular speaker’s portion within a video by simply clicking on their face in the timeline. Moreover, viewers can search for video content by a specific speaker by simply typing their name in the search bar.
- Search by key topics: Videos under the same topic category can be collectively bundled and stored in playlists within VIDIZMO, which further improves its search effectiveness. A simple search would reveal results for all videos categorized under a particular topic, either individually or within an organized playlist.
- Search by words that appear in a video: VIDIZMO utilizes optical character recognition (OCR) technology to detect and index text that appears in a video and converts it to a machine-readable text file for your use. This means you can extract meaningful data from the text in videos, such as in the case of learning or training material presented in videos, which can then be indexed into easily searchable text.
- Search by video timestamp: VIDIZMO allows search based on a given time code or segment within a video. For instance, you can simply search a time interval in a video to refine search from that point onwards. This is extremely beneficial for video sharing between peers so the shared video starts and ends at the given timestamp, directing the viewer to a specific portion of the videos and saving them time.
- Search by language: VIDIZMO is a multi-lingual software available in a variety of languages, including English, Spanish, French, and Arabic. For search in a particular language that isn’t English, VIDIZMO enables indexing of all translated video transcripts and caption files so the user can search for terms within the video in their preferred language.
- Search across video library using traditional & custom metadata: VIDIZMO allows a thorough and detailed search of all videos in the platform using metadata such as video title, description, tags, hashtags, author or even by custom attributes such as publishing date, video type, live event date, department, company, location or any other required identifiers.
- Search videos using mobile app: VIDIZMO mobile application, available on Android and iOS, has both traditional and smart search capabilities that enable you to search using any of the categories mentioned above, straight from VIDIZMO's mobile application.
- Search within videos from integrated business applications: VIDIZMO allows you to do in-video search from videos obtained via integration with third-party applications. For instance, VIDIZMO can enable traditional or smart search for meeting videos automatically ingested from integrated video conferencing software such as Skype, Go-To-Meeting or WebEx.
With all the above-mentioned search capabilities, VIDIZMO provides a search inside video solution that stands unparalleled to what most enterprise video platform vendors in the market can offer. For smart video search capabilities such as ASR, OCR and facial detection, VIDIZMO utilizes artificial intelligence technologies developed by industry giants such as Microsoft Azure and IBM Watson to deliver smart video search capabilities within VIDIZMO’s enterprise video platform.